The Sentencing Project put out a new analysis this morning finding little support for a so-called “Ferguson effect” on crime in St. Louis, Missouri. The “Ferguson effect” describes a conjecture by some commentators that rising crime rates in some urban areas in recent months are the result of widespread protests against police misconduct and calls for reform.
Those demonstrations spread across the nation in response to a stream of highly publicized killings of unarmed black men and boys by police, starting with the death of 18-year-old Michael Brown in Ferguson, Missouri, last August.
Since Ferguson is part of the St. Louis greater metropolitan area, one could expect that if such an effect exists anywhere, it would be in St. Louis. To determine whether such an effect exists, criminologist Richard Rosenfeld analyzed trend data on homicides, violent crimes, and property crime in St. Louis in recent years.
The report, Was There a “Ferguson Effect” on Crime in St. Louis?, finds that the homicide count in St. Louis was higher in 2014 than in 2013 for most of the year, but the gap between the two years began to increase two months before the events in Ferguson. “We can conclude with reasonable certainty that the events in Ferguson were not responsible for the steep rise in homicide in St. Louis,” the report states.
A four-month increase in crime is not a crime wave, and while the homicide rate in St. Louis rose in 2014, overall it has declined significantly since the 1990s. This report underscores the importance of policies based on rational analysis of crime trends that appreciate the complexity of factors that produce crime.
Was There a “Ferguson Effect” on Crime in St. Louis?
Richard Rosenfeld, University of Missouri – St. Louis
Does widespread and heavily publicized protest activity directed at alleged police misconduct result in higher crime rates? That is the view of some commentators, who attribute recent crime increases to what has been termed the “Ferguson effect.” The idea is that heightened criticism of the police has demoralized, angered, frustrated, or otherwise caused police officers to refrain from vigorous enforcement activity, resulting in more crime (e.g., Mac Donald 2015).
Is there evidence that crime rates are, in fact, increasing around the country? It depends on where you look. In some cities, some types of crime are up over last year and other types are not, while other cities have not recorded increases in any major crime category. Unfortunately, we cannot rely on the nation’s leading crime data systems for a comprehensive view of recent crime changes. The Bureau of Justice Statistics’ National Crime Victimization Survey does not produce city-level crime data. The FBI’s Uniform Crime Reports does provide crime data for individual cities, but the data are not timely enough for monitoring crime changes over the past several months. So, everybody’s data is anecdotal, a cherry picker’s delight. If you want to tell a story of crime increases, you can. If not, just pick from a different tree.
ANALYZING A POSSIBLE “FERGUSON EFFECT”
So, how might we evaluate the so-called Ferguson effect on crime? One way is to go to the source, or close to it, and examine crime changes in St. Louis, located just a few miles from Ferguson, in the months before and after Michael Brown was killed. A superficial glance at St. Louis crime statistics might lend some credence to the view that crime increased in response to the events in Ferguson. After declining or holding steady for several years, the number of violent crimes in St. Louis rose by 5.3% in 2014 over the previous year. The increase in homicide was especially pronounced. In 2014, the police department recorded 159 homicides, a 32.5% increase over the 2013 total of 120. Local officials as well as national commentators attributed the crime increase in St. Louis to the police shooting in Ferguson that led to protest demonstrations in St. Louis as well as Ferguson.
To determine with certainty whether the Ferguson events caused crime increases in St. Louis (or elsewhere), we would have to know what would have happened to crime had the Ferguson events not happened. Proving the counterfactual – whether y would not have occurred had x not occurred – is challenging, to say the least. In this case, however, we can begin by asking whether the timing of the Ferguson effect hypothesis corresponds with observed crime changes in St. Louis. Obviously, the Ferguson events or subsequent police shootings could not have produced crime increases in St. Louis if those increases took place before Michael Brown was killed. If the increases occurred afterward, that would not prove they were caused by the controversial police shootings, but it is a necessary condition for drawing causal inferences.
Figure 1 displays month-by-month homicide frequencies in St. Louis in 2013 and 2014. The homicide count was higher in 2014 than in 2013 throughout most of the year, but it is not readily apparent from the figure whether the gap grew after Michael Brown was killed in early August of 2014. To establish whether the police shooting in Ferguson and resulting protests may have triggered an increase in homicide, the month-by-month ratio of homicides in 2014 to homicides in 2013 is shown in Figure 2. Here we see that homicides were far more frequent in the early months of 2014 compared with the same period in 2013. The ratio then decreased but began to rise again in June, two months before Michael Brown was killed, and at no time in the following months did it exceed the levels reached earlier in the year. These results do not support the belief that the St. Louis homicide increase in 2014 was attributable to the Ferguson events.
OTHER VIOLENT CRIME RATES
The picture differs somewhat for other violent crimes. Figure 3 displays the monthly ratio of violent crimes in 2014 to violent crimes in 2013. We see that the ratio grew early in the year, decreased for two months, and then increased again through the end of the year. But that increase began in May, well before the police shooting in Ferguson. The rate of increase in the violent crime ratio, however, does appear to have accelerated beginning in September. The timing of these changes offers mixed support, at best, for the Ferguson effect hypothesis.
PROPERTY CRIME RATES
Figure 4 displays the ratio of property crimes in 2014 over 2013. The monthly change in the property crime ratio offers the strongest evidence in support of the hypothesis that the Ferguson events led to crime increases in St. Louis, at least with respect to timing. The property crime ratio was fairly flat until August, when it began to increase steadily through the remainder of the year. Before August, St. Louis averaged about 15% fewer property crimes in 2014 than in 2013. From August through the end of the year, the city averaged 8% more property crimes than during the same period the year before. By December, the number of property crimes in 2014 exceeded the number in 2013 by 27%. If there was a Ferguson effect on crime in St. Louis, it was most pronounced in the growth of property crimes.
CRIME DATA SUMMARY
Even limiting an analysis just to the timing of these monthly crime changes provides only partial support for the contention that the police shooting in Ferguson and ensuing protest activity spurred crime increases in St. Louis. The increase in homicide in 2014 predated Michael Brown’s killing on August 9. Violent crimes exhibit much the same pattern, although the rate of increase in violent crimes over 2013 accelerated after August. Only the timing of the change in property crimes is fully consistent with a Ferguson effect. But temporal consistency is not a sufficient condition to establish substantive proof. Many factors may have contributed to a rise in property crimes, and perhaps also violent crimes, in St. Louis after August of 2014, and we cannot conduct an experiment to discover whether crime would have increased had Michael Brown not been killed in a controversial police shooting. But we can conclude with reasonable certainty that the events in Ferguson were not responsible for the steep rise in homicide in St. Louis.
Whatever their cause, double-digit homicide increases in St. Louis and other cities during the past several months should not be discounted as unimportant or as mere “random fluctuations” in crime statistics – not when so many lives are at stake. But neither should the recent increases be read as a new crime wave, at least not yet. Even with a 33% increase over the previous year, the St. Louis homicide rate in 2014 remained well below the peak levels reached in the early 1990s (see Figure 5). We should watch the new increases carefully and encourage local police departments to apply evidence-based best practices in response. In the absence of credible and comprehensive evidence, sounding alarm bells over a “Ferguson effect” or any other putative cause will not help.